A novel mathematical model for modeling viscosity and temperature relationship for dead oils

Area of Science:

• Petroleum Engineering
• Fluid Dynamics
• Chemical Engineering

Background:

• Accurate viscosity prediction is vital for subsurface and surface transport, heat transfer, and pipeline design.
• Existing viscosity models for hydrocarbon fluids often lack precision or are limited in scope.
• Forecasting dead oil viscosity accurately remains a significant challenge in the industry.

Purpose of the Study:

• To develop and validate a mathematical algorithm for accurate estimation of viscosity in hydrocarbon fluids.
• To establish a reliable relationship between fluid viscosity and temperature using non-linear regression.
• To assess the algorithm's performance across a range of crude oil types and varying data availability.

Main Methods:

• A robust non-linear regression technique was employed to model the relationship between viscosity and temperature.
• The algorithm was applied to a dataset of 243 viscosity data points from Iranian crude oil samples (extra-heavy to light).
• Model performance was evaluated using metrics such as maximum absolute error, relative error, and coefficient of determination.

Main Results:

• The algorithm accurately estimated viscosity for diverse crude oil samples, with values ranging from 0.29 cp to 5328.74 cp.
• The highest errors (1.25 cp absolute, 6.04% relative) were observed for a fluid with API gravity of 12.92.
• Effective models were trained using less than 30% of the available data, showing high precision.
• Models achieved a near-unity coefficient of determination on testing data, indicating strong empirical data reflection.

Conclusions:

• The developed mathematical algorithm provides a precise and reliable method for estimating hydrocarbon fluid viscosity.
• The algorithm's effectiveness is demonstrated even with limited training data, making it suitable for practical applications.
• The study confirms the algorithm's proficiency in accurately reflecting empirical viscosity data across various crude oil types.